The European Union has introduced a new regulation on artificial intelligence that requires employers to inform workers and their representatives before using high-risk AI systems in the workplace. The regulation, specifically article 26, paragraph 7, mandates that this information be given in accordance with Union law, national law, and established practices. The term "deployers" refers to those who use AI systems "under their own authority," typically employers using software designed by a publisher. The regulation primarily applies to employees in French companies, though its scope may extend to other EU member states. High-risk AI systems, as defined in annex III of the regulation, include those used for recruitment, decisions affecting employment conditions, promotion or dismissal, assigning tasks based on personal criteria, and monitoring employee performance and behavior. The regulation allows for an exclusion if the system does not pose a significant risk, such as when it performs a narrow procedural task. However, systems that create employee profiles are always considered high risk. The obligation for informing workers will take effect on December 2, 2027, a date postponed from August 2, 2026. The legal basis for the regulation is found in articles 16 and 114 of the Treaty on the Functioning of the European Union (TFUE). Article 114 relates to the internal market and allows harmonization measures aimed at establishing and functioning of the internal market, though it excludes tax provisions, free movement of persons, and rights and interests of employees. The regulation's information obligation is considered "accessory" to the main objective of protecting fundamental rights, a characterization used by the Court of Justice of the European Union in 2025. The regulation does not specify the form of the information, only that it must precede the first use of the AI tool. Employers must be able to justify compliance, and French law provides a benchmark for informing employees of the notice of absence. The regulation does not apply to systems already marketed or in use before December 2, 2027, unless they undergo significant design changes. Non-compliance can result in administrative fines up to 15 million euros or 3% of annual global turnover, whichever is higher. France is considering entrusting the CNIL with oversight of AI systems used for employment and personnel management, awaiting approval by the National Assembly.